Support Vector Regression Based Reconstruction Algorithm for Concentrations of Ternary Solution

WEI Guo,Xin Liu,Jinwei Sun,SUN Sheng-he
DOI: https://doi.org/10.3321/j.issn:0372-2112.2008.04.016
2008-01-01
Abstract:Ternary solution with NaCl and sucrose is widely employed in the osmotic dehydration process of food.In this paper,a novel multifunctional sensor was adopted to online sense temperature,ultrasonic velocity and electrical conductivity of ternary solution.With these three measurable parameters as basis,this study laid emphasis on the support vector regression(SVR)method used to implement multifunctional sensor signal reconstruction,where the concentrations of two components in ternary solution with NaCl and sucrose can be simultaneously estimated.Support vector machine(SVM)is a new machine learning method based on structural risk minimization,which is well adapted to small sample size problem of calibration data and can efficiently restrain over-fitting and improve the generalization capability.The experimental results show that the mean absolute errors of the reconstructed concentrations of NaCl and sucrose for test data set are 0.00615 and 0.00369mol/kg,respectively.It could demonstrate the high reliability and high accuracy of the proposed reconstruction algorithm and verify the feasibility.
What problem does this paper attempt to address?